Changelog

3.5.3 

Fixed

Fixed G5k firewall deployment

3.5.2 

Fixed

Fixed IPv6 enabling in G5K interface handling.

3.5.1 

Removed

Removed pytz dependency

3.5.0 

Added

Added __g5k_ipv6_address__ extra information for grid5000 nodes.
Documentation
Added --ansible_forks option to set ansible forks (the number of concurrent tasks run on hosts) number which can speed up workflow deployment and execution.
Documentation

Changed

Improved infrastructure deployment (layers-services or `` deploy`` command) by removing redundant steps

3.4.0 

Changed

e2clab ssh can now be run asynchronously while another e2clab command is running once an infrastructure is deployed.
Now uses a .e2c_state dir to store experiment state

Added

Added options --destroy-on-fail for deploy to automatically free resources in failed experiments.
Documentation
Added option --destroy for the destroy command to destroy deployed network emulations.
Documentation
Added option --workflow_env for the init command to initialize a project using worflow_env
Documentation
Added option --pause/-p for the deploy command to wait for user input before finalizing experiment
Documentation
init command now initiates SCENARIO_DIR with template files.
Documentation
Added ipv6 option for G5K testbed enabling public IPv6 addresses for nodes.
Documentation

3.3.1 

Fixed

Major backup bug for monitoring kwollect manager

3.3.0 

Added

Added support for EnOSlib 10
Added options --destroy for the finalize and deploy commands, allowing for automatic freeing of computing resources.
Documentation
Added a ‘Kwollect’ monitoring manager, allowing users to use and pull metrics from the Grid’5000 API
Documentation

Changed

e2clab workfow SCENARIO_DIR launch --app_conf APP_CONf_LIST no longer runs finalize after each app_conf. Use --finalize tag for this.
Documentation

Fixed

Loggers outputing to previous experiments files when repeating deploy experiments (#46)
Loggers not outputing to logfiles when running workflows incrementally (#44)
workflow_env syntax checking failing when app_conf names are digits (#39)

3.2.2 

Fixed

Fixed import bug that only allowed experiments containing Iotlab resources (#32)

3.2.1 

Fixed

Fixed e2clab ssh SCENARIO_DIR not understanding environment variables (#30)

3.2.0 

Added

Configure e2clab with environment variables like E2C_SCENARIO_DIR or E2C_ARTIFACTS_DIR to override command line arguments and default variables.
Documentation
Configure e2clab with a .e2c_env file which can be used to automatically set environment variables during command run.
Documentation
New commands:
- e2clab ssh: Prompts the user for a host to ssh to. Documentation
- e2clab edit: Command to quickly open an editor to scenario’s configuration file. Documentation
- e2clab init: Command to quickly create .e2c_env file for your experiment. Documentation
- e2clab get-output-dir: Print latest experiment dir to stdout.
  Documentation

Documentation

Enabling tab completion on your shell

3.1.0 

Changed

[BREAKING] Service API parameter syntax change for user-defined services in register_service method.
Documentation

3.0.0 

Added

New commands:
- e2clab services: manage user-defined services without having to copy files around. Documentation
- e2clab destroy: frees testbed resources with a single command. Documentation
- e2clab check-configuration: checks configuration file syntax.Documentation
Add suppport for EnOSlib 9
New worflow_env.yaml file to improve --app_conf capabilities. Documentation
- Allows to chain run differrent configurations of workflow.yaml.
New e2clab.log and e2clab.err logging files in experiment’s output folders.
New CLI command-line options, documentation:
- --debug parameter to allow debug-level logging.
- --mute_enoslib mute inherited enoslib logging.
- --mute_ansible mute ansible spinning callback in terminal.

Changed

[BREAKING] Syntax verification for yaml configuration files enforced by jsonschemas. Documentation
- Some syntaxic changes (e.g. depends_on attribute in workflow.yaml must now be an array)
- Error logs should be verbose enough to fix configuration files.
- Refer to the schemas in the documentation as a baseline reference.
Improved logging in terminal
- e2clab.log and e2clab.err log files in experiment result dir.
- --debug parameter to allow debug-level logging.
- --mute_enoslib mute inherited enoslib logging.
- --mute_ansible mute ansible spinning callback in terminal.
[BREAKING] Changed optimize workflow. Documentattion

Removed

[BREAKING] Retired optimize command. Instead directly use API, see documentation

Documentation

2.1.0 

Added

Provenance Data Manager: allows users to capture provenance data of Edge-to-Cloud workflows at runtime

Documentation

2.0.0 

Added

New Providers: Chameleon Cloud and CHI@Edge
Multi-platform workflow deployment
TPG monitoring: Telegraf + Promotheus + Grafana
FIT IoT LAB power consumption monitoring
New CLI commands such as check-testbeds and new parameters for deploy (scenarios_name and app_conf) and workflow (app_conf and finalize_wf)
Add support for EnOSlib 8.1.5
Reusable User-Defined-Services (+10 services)

Changed

New definition of layers_services.yaml for the following attributes environment ( to support multiple testbeds), monitoring (switch between Dstat, TIG, and TPG), services (link to different testbeds)
NetemHTB for network emulation